Mailing-List: contact dev-help@tomcat.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Developers List" <dev@tomcat.apache.org>
Received-SPF: softfail (athena.apache.org: transitioning domain of
 brian@pingtoo.com does not designate 217.154.193.216 as permitted sender)
Message-ID: <50AE7C97.3090206@pingtoo.com>
Date: Thu, 22 Nov 2012 19:27:19 +0000
From: Brian Burch <brian@pingtoo.com>
User-Agent: Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/17.0 Thunderbird/17.0
MIME-Version: 1.0
To: dev@tomcat.apache.org
Subject: Re: [Bug 54190] TestNonLoginAndBasicAuthenticator does not test
 session
 timeout properly
References: <bug-54190-78@https.issues.apache.org/bugzilla/>
 <bug-54190-78-xcQRK2BLgm@https.issues.apache.org/bugzilla/>
 <50AE61AE.70703@pingtoo.com> <50AE7167.7010309@apache.org>
In-Reply-To: <50AE7167.7010309@apache.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 22/11/12 18:39, Mark Thomas wrote:
> On 22/11/2012 17:32, Brian Burch wrote:
>> On 22/11/12 16:46, bugzilla@apache.org wrote:
>>> https://issues.apache.org/bugzilla/show_bug.cgi?id=54190
>>>
>>> --- Comment #3 from Mark Thomas <markt@apache.org> ---
>>> Your logic makes sense to me so my preference would be some more
>>> comments.
>>
>> I will think about our discussion and try to improve the comments and so
>> hopefully avoid confusing future readers. I will submit a new
>> self-contained patch and obsolete the current one.
>>
>>
>> Sorry to drift slightly (but not completely) off-topic. Could you give
>> me your off-the-cuff opinion on this:
>>
>> Context.setSessionTimeout(int timeoutInMinutes) obliges tests that need
>> to explore session timeout behaviour to hang the test process for at
>> least 60 seconds, although most of that delay would be unnecessary.
>>
>> I don't actually know how often the Manager (is that the right
>> component?) scans to expire Sessions, but it must do it more frequently
>> than once a minute.
>
> It is the manager. It scans (goes to look up frequency) every 60 seconds
> by default (10s for background process * 6 for processExpiresFrequency)
>
>> I realise the method signature is part of a public api, but do you have
>> a view on adding an alternative method to
>> org.apache.catalina.core.StandardContext so that unit tests could set a
>> session timeout in seconds?
>
> I'm neutral. The delay doesn't really bother me. If you do go this way,
> you'll need to modify processExpiresFrequency for the Manager as well so
> checks happen more often.

Perfect answer, thanks! If the Manager by default is scanning every 60 
seconds, then it makes sense to live with a session timeout of 1 minute 
and a test case wait of about 70 seconds. That is what I am doing at the 
moment, and there isn't much point being clever unless a test suite is 
going to call for more than a few timeouts.

> Mark
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
> For additional commands, e-mail: dev-help@tomcat.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org